Conference
Informed POMDP: Leveraging Additional Information in Model-Based RL
| Τίτλος: | Informed POMDP: Leveraging Additional Information in Model-Based RL |
|---|---|
| Συγγραφείς: | Lambrechts, Gaspard, Bolland, Adrien, Ernst, Damien |
| Πηγή: | Reinforcement Learning Journal (2024-08); Reinforcement Learning Conference, Amherst, United States - Massachusetts [US-MA], August 9th, 2024 |
| Έτος έκδοσης: | 2024 |
| Θεματικοί όροι: | Computer Science - Learning, Engineering, computing & technology, Computer science, Ingénierie, informatique & technologie, Sciences informatiques |
| Περιγραφή: | In this work, we generalize the problem of learning through interaction in a POMDP by accounting for eventual additional information available at training time. First, we introduce the informed POMDP, a new learning paradigm offering a clear distinction between the information at training and the observation at execution. Next, we propose an objective that leverages this information for learning a sufficient statistic of the history for the optimal control. We then adapt this informed objective to learn a world model able to sample latent trajectories. Finally, we empirically show a learning speed improvement in several environments using this informed world model in the Dreamer algorithm. These results and the simplicity of the proposed adaptation advocate for a systematic consideration of eventual additional information when learning in a POMDP using model-based RL. |
| Τύπος εγγράφου: | conference paper http://purl.org/coar/resource_type/c_5794 conferenceObject peer reviewed |
| Γλώσσα: | English |
| Relation: | https://arxiv.org/abs/2306.11488; urn:issn:2996-8569; urn:issn:2996-8577 |
| Σύνδεσμος πρόσβασης: | https://orbi.uliege.be/handle/2268/304369 |
| Rights: | open access http://purl.org/coar/access_right/c_abf2 info:eu-repo/semantics/openAccess |
| Αριθμός Καταχώρησης: | edsorb.304369 |
| Βάση Δεδομένων: | ORBi |
| Η περιγραφή δεν είναι διαθέσιμη |